Your AI Agent Just Committed a Federal Crime — Inside the Ruling That Could Kill Agentic Commerce
A federal judge in San Francisco just ruled that your AI shopping assistant may be breaking the same law used to prosecute hackers — even when you explicitly told it to act on your behalf. The March 2026 Amazon v. Perplexity decision draws a line that could reshape the entire AI agent industry: user permission is not platform permission.
The implications extend far beyond one company's browser. As 17,000+ autonomous agents execute millions of daily transactions across Web2 and Web3, this ruling forces a fundamental question: who actually authorizes an AI agent to act — the person who deployed it, or the platform it touches?